1. Field of the Invention
The present invention relates to a method of sequence determination for nucleic acid such as DNA (deoxyribonucleic acid), and more particularly, it relates to a method of sequence determination for nucleic acid characterized in pretreatment carried out for removing noise before determining the base sequence from data obtained by electrophoresis.
2. Description of the Prior Art
In a method of electrophoresing a fragment sample of nucleic acid and determining the base sequence of the nucleic acid on the basis of detected data, the data detected by electrophoresis is a peak signal corresponding to the fragment samples of the nucleic acid. Since the peak signal includes noise components, waveform shaping is performed on the peak signal by pretreatment, for thereafter determining the base sequence on the basis of the peak signal.
The pretreatment is generally collectively performed on all detected data obtained by electrophoresis for thereafter performing a sequence determination (base calling) with treated waveforms.
The pretreatment for waveform shaping includes waveform shaping by Fourier transformation represented by FFT (fast Fourier transformation). In this pretreatment, filtering is performed when performing Fourier transformation on data of a peak signal detected by electrophoresis for thereafter returning the data to the peak signal by inverse Fourier transformation. A filter function employed in this filtering is adapted to remove signals having frequencies shorter than a DNA peak appearance interval, in order to mainly remove noise which is a high-frequency component.
{circle around (1)} While FFT must be performed on 2n data, the number of the data, varying with the migration time or the sampling frequency, is not constant.
{circle around (2)} The filter function for noise removal is set for mainly removing noise which is a high-frequency component, i.e. a signal having a frequency shorter than the DNA peak appearance interval. Therefore, the migration speed serves as an important parameter for the filter function. However, the migration speed gradually changes during migration, and hence noise filtering cannot be performed with the same parameter over the overall data area.